Been using EAC for ripping for some time now (LAME alt preset standard), and I notice that many of my rips come out at 99%, but EAC also reports no errors. Is this normal? These are perfectly clean CDs btw.
So should I be concerned? I certainly can't hear anything, but my ears nor my audio knowledge is no where near as keen as most around here.
I must add I have not messed with drive offsets yet, since everything seems to sound okay to me, and I was under the impression that offsets where really for exact copies where even the original gaps are desired. (No need for offset discussion if that's not my 1% missing)
Thanks as always.
No you shouldn't be worried. This is from the excellent Coaster Factory Exact Audio Copy Tutorial (http://users.pandora.be/satcp/tutorials.htm):
Some explanation of the information displayed in the Status and Error Messages log. Peak level is the loudest peak in the extracted track. Though the example screenshot has all peak levels at 100%, the maximum volume, this is not necessary true for all CDs. Some CDs (especially older ones) are recorded a lot softer. Thus if the Peak level shows a low percentage for certain tracks, don't worry. This does not mean that Exact Audio Copy was not able to extract the tracks 100% correct!!
While Peak level tells nothing about the quality of extraction, Track quality does. A Track quality of 100% obviously means that the track was extracted 100% correct. But here's where some people make mistakes; sometimes EAC rereads certain audio sectors multiple times to get accurate extraction results. For every reread EAC does, the Track quality decreases, but this does not mean that the extraction is less accurate. It is possible to have a bit-by-bit perfect copy of a track, while Track quality is lower than 100%. As long as Exact Audio Copy does not report any errors in the Status and Error Messages log, the extracted files are bit-by-bit perfect copies of the original. Track quality should be interpreted as the physical quality of the CD and not of the extracted data. A CD with some scratches or dirty fingers on will certainly cause rereads in EAC and thus a Track quality lower than 100%, but still the extracted tracks may be perfect. Thus if the log says Copy OK for a track that means it's extracted perfect - no matter of the Track quality. So, I hope that made things more clear as many people are confused by the Track quality.
Excellent. Thank you for the quick response.
I read their tutorial (okay, parts) some time ago, but I don't recall reading that particular bit of information. Explains it perfectly.
Thanks again.
Speaking of track quality, how many CDs have tracks with quality less than 100%? I am just curious. I have ripped many brand new, un-scratched, out-of-the-sealed-box CDS with NO fingerprints and most of them have at least one track with quailty of 99.99%. The number of such tracks varies from 1 to as much as 7 per disc. The average is about two tracks per disc. I am worried about the physical condition of the discs. Since they are brand new, I expect them to be perfect, but as you can see, it is not so.
How many people have seen this happening while you rip discs? Please let me know, thanks.
Speaking of track quality, how many CDs have tracks with quality less than 100%? I am just curious. I have ripped many brand new, un-scratched, out-of-the-sealed-box CDS with NO fingerprints and most of them have at least one track with quailty of 99.99%. The number of such tracks varies from 1 to as much as 7 per disc. The average is about two tracks per disc. I am worried about the physical condition of the discs. Since they are brand new, I expect them to be perfect, but as you can see, it is not so.
A few, no question about it, but this is to be expected with an aluminum reflecting layer. The level of control over the layer of aluminum is not total even at the best plants due in part to the physical properties of aluminum (and the laquer layer coating the top of the disc), and so it's not that suprising that EAC would have to go over a very short segment more than once even on a brand new CD.
The quote above of course explains why "track quality" isn't necessarily the quality of the ripped track.
Speaking of track quality, how many CDs have tracks with quality less than 100%? I am just curious. I have ripped many brand new, un-scratched, out-of-the-sealed-box CDS with NO fingerprints and most of them have at least one track with quailty of 99.99%. The number of such tracks varies from 1 to as much as 7 per disc. The average is about two tracks per disc. I am worried about the physical condition of the discs. Since they are brand new, I expect them to be perfect, but as you can see, it is not so.
How many people have seen this happening while you rip discs? Please let me know, thanks.
I find this occurence of less than perfect tracks on new CD (or CD in otherwise excellent shape) on a regular basis.
As long as EAC does not report any errors, I don't even bother to check the WAV.
Asf it has been pointed out, less than 100% means EAC had to re-read a few times, but most of the reads were consistent, hence EAC assumes it read the segment correctly.
If you watch the ripping process, you'll see few red boxes 'lights up' - indication that EAC had to re-read. Normally it happens really quickly. Then the quality will be less than 100%.
I have noticed that carefully cleaning the media before ripping decreases the occurence of 'less than 100%'.
On the older media - seems like some CDs do not age as well as the others. In particular, I always has some problem with 'Errato' label (classical/20 century). (they should have choosen different name :-) )
Don't forget that reading a CD is an active, mechanical process.
Transient errors can arise in the process that EAC deals with superbly by its re-read technique. It may be your CD-ROM is partly to blame for some errors with otherwise perfect (new) CD's
Speaking of track quality, how many CDs have tracks with quality less than 100%? I am just curious. I have ripped many brand new, un-scratched, out-of-the-sealed-box CDS with NO fingerprints and most of them have at least one track with quailty of 99.99%. The number of such tracks varies from 1 to as much as 7 per disc. The average is about two tracks per disc. I am worried about the physical condition of the discs. Since they are brand new, I expect them to be perfect, but as you can see, it is not so.
How many people have seen this happening while you rip discs?
I have seen this so many times I cannot count it! Especially the two tracks per disc rings a loud bell, that happens all the time.
The strangest thing though is I've bought used CD's that look like they would be an un-rippable nightmare that have 100% quality.
If you're ripping and encoding with EAC, I've found that using the Compression Queue Control Center has dramatically reduced the number of 99.9% results I see. The location has recently changed with the latest release of EAC, but I believe the keypress of ctrl-Q should work in all version.
Set the control center to "make all compression tasks sleep", and release it after you've ripped your cd (or multiple cds to encode at night, for example). YMMV, but I'm very pleased with what I've seen (as I know it's more of a "peace of mind issue, not one I can hear").
If there is missing printing on the top of the cd, where you can see right through the cd, you can never get a copy ok. Is there an easy way to copy those kind of tracks?
I get 99% very often, but the reread always occur at track changes, and especially at the end of the CD itself.
There can be several explanations :
-The mastering was done track-at-once and not disc-at-once.
-The track change is a sensitive part, prone to errors (in a software way).
-There is something spooky occuring at tracks changes (in a hardware way), that explains why the tracks are visible to the naked eye on the CD.
-CD ROM drives are bugged.
-EAC is bugged.
The third hypothesis needs an explanation anyway. What is it that changes the color of the CD between each track, and makes it look like vinyl tracks ? I really don't know, and I would be curious to know.
The fourth hypothesis is right with my memorex DVDMaxx1648, and was probably right with my older drives, since I always got problems at the end of CDs, even without offset correction. My drive silences a random number of sectors at the end of last track. When there is an analog noise, or a dither signal in the silence after the last track, this leads to mismatching CRCs, until the drive cuts the same amount of audio two times in a row.
The fifth hypothesis was reported by Andre Wiethoff long ago. He talked about the synch between track process triggering error correction by mistake. But this can have been corrected since then.
I get 99% very often, but the reread always occur at track changes,...
My experience is nearly the same as Pio's. On average not 100,0%, and the reason is not the physical quality of the CD, appears 1-2 times per CD and always at the end of tracks, but this never causes an error, CRC is OK and it's always only the first row of the red lights. It seems to me that this is a bug in EAC because I got this with several drives. Maybe this has something to do with the option "synchronize between tracks" ?
In principle, CRC=OK is a
very very secure thing. I've only had one single case so far when the CRCs of two wavs from the same track were OK and the wavs were different (out of a thousand or so); and there was much error correction because of scratches.
-There is something spooky occuring at tracks changes (in a hardware way), that explains why the tracks are visible to the naked eye on the CD.
The third hypothesis needs an explanation anyway. What is it that changes the color of the CD between each track, and makes it look like vinyl tracks ? I really don't know, and I would be curious to know.
I always thought this are the gaps on the CD, visible because there are only null samples. And if the gap is longer, you can see that this 'spooky' area is bigger.
You're right, a given amount of perfectly null samples will produce a regular pattern, diffracting light differently.
In this case, CDs with dither noise between tracks would not have any visible marks.
Maybe I shouyld compute the DSV of null samples, and see if they behave like weak sectors...
You're right, a given amount of perfectly null samples will produce a regular pattern, diffracting light differently.
In this case, CDs with dither noise between tracks would not have any visible marks.
Maybe I shouyld compute the DSV of null samples, and see if they behave like weak sectors...
What is DSV ?
During the last days I was wondering about this question. I analysed a CD that hasn't any gaps (according to EAC), but the spooky areas were still visible. I did the same with a CD that has gaps between all tracks and all gaps are 0% silence, according to EAC, with the same result. On my self-made CDs (I never write any gaps) the tracks are visible, too. I decided to make a
very very loud CD with no fade in and fade outs to see if they are the reason (because of a higher percentage of 0s), but the result wasn't unambiguous.
So it seems that we are wrong.
Any ideas ?
I get 99% very often, but the reread always occur at track changes, and especially at the end of the CD itself.
I have been experiencing the same thing as you've explained for the last few days since I found a stack of un-ripped CD's.
What I don't understand is how some tracks can get 99% or lower quality when leaving the drive speed to Actual, versus changing to to say 48X and then getting a quality of 100%.
I noticed some behaviour with my drive speed set at Actual when ripping a very scratched and scuffed used CD as EAC would stop at points in a track for several minutes without any what I would call activity, and then slowly begin again only to repeat the process. The extraction was at 40 minutes and wasn't even half done yet. When I decided to give up out of fear of damaging the drive and then configured EAC to use the drive maximum speed at 48X which for some reason isn't always selectable it did the secure mode error correction as if it were just a minor problem spot and finished the disc in about 10 minutes.
The condition the disc was in made me think that I would have to resort to using CDex. However reconfiguring EAC worked, both Test and Copy CRC's matched.
What I don't understand is how some tracks can get 99% or lower quality when leaving the drive speed to Actual, versus changing to to say 48X and then getting a quality of 100%.
I noticed some behaviour with my drive speed set at Actual when ripping a very scratched and scuffed used CD as EAC would stop at points in a track for several minutes without any what I would call activity, and then slowly begin again only to repeat the process. The extraction was at 40 minutes and wasn't even half done yet. When I decided to give up out of fear of damaging the drive and then configured EAC to use the drive maximum speed at 48X which for some reason isn't always selectable it did the secure mode error correction as if it were just a minor problem spot and finished the disc in about 10 minutes.
Are you really sure that this extraction was exact? Was the CRC OK?
Did you extract several times in the same way and the files were identical ?
Did you verify (compare) your extraction with an other extraction from an other drive or with the recording from the digital out of an external CD-player ?
Did you use accuraterip and the tracks were accurately ripped ?
Even if the CRC was OK, the extraction still can be wrong. I once had this case (only one times). To make me believe what you said, you have to do all of the above.
Because from all of my experience with EAC, such a behavior means that EAC doesn't work correctly with the speed set to 48x. But I've no experience with that, there's always only "actual".
Are you really sure that this extraction was exact? Was the CRC OK?
Did you extract several times in the same way and the files were identical ?
Did you verify (compare) your extraction with an other extraction from an other drive or with the recording from the digital out of an external CD-player ?
Did you use accuraterip and the tracks were accurately ripped ?
Even if the CRC was OK, the extraction still can be wrong. I once had this case (only one times). To make me believe what you said, you have to do all of the above.
Because from all of my experience with EAC, such a behavior means that EAC doesn't work correctly with the speed set to 48x. But I've no experience with that, there's always only "actual".
Read the last few words of my post, both Test and Copy CRC's match in the EAC window. In the EAC window it calls the extracted CRC "Read", in the generated ripping log it calls the extracted CRC "Copy". Both CRC's match, they matched with two different DVD drives; JLMS DVD-ROM XJ-HD166 (can have speeds set) and TEAC DVD+RW DV-W58E (actual speed must be selected or extraction is completely unreliable).
I have no ability to do digital out with an external/home stereo type CD player, since mine only has two RCA output jacks (analog).
I know that matching CRC's are no gaurantee of a flawless and error-free extraction, after all I've been dealing with two drives that cache audio, I've noticed before when errors go undetected even when the CRC's match.
I do not use AccurateRip, I'll have to look into it and hopefully the CD in question is in the database.
I have no ideal if EAC is or isn't working correctly or not and have no comment on that, I am not making any statements about EAC's quality or stability. However, what I do think is happening is the drive(s) caching audio that is causing the majority of the problems, e.g.; the insanely lengthy slowdown when EAC has to clear the cache or whatever it is doing.
The JLMS reader drive on bad discs has performed "better" by selecting a set speed versus actual, it doesn't necessarily have to be the maximum of 48x, it could be 16x for instance, plain and simple it doesn't get stuck in a certain bad spot of a disc when doing this. The TEAC writer drive can only be set to 40x, but I already know it is unreliable when selecting a set speed therefore I only use actual with it, also Feurio's test state that speeds cannot be set with the TEAC writer drive, however it doesn't get stuck for long periods on bad disc -- but I don't want to stress it or even use it for audio extraction since it is the only writer I have and need for backups.
> Maybe I shouyld compute the DSV of null samples, and see if they behave like weak sectors...
That would be a terrible flaw in the EFM code, and luckily it's not the case.
> Maybe I shouyld compute the DSV of null samples, and see if they behave like weak sectors...
That would be a terrible flaw in the EFM code, and luckily it's not the case.
Flaws in EFM are already being implemented in new drives to remove they ability to copy certain (lice Safedisc) copy-damaged titles.
> Flaws in EFM are already being implemented in new drives to remove they ability to
> copy certain (lice Safedisc) copy-damaged titles.
If you have any evidence of this then post them along with the technical explanation,
otherwise I will consider that you're just repeating the same false rumour we've seen
for years.
Well, maybe that was just that friend's Plextor I can recall...
after fimware upgrade it stopped writing EFMs correctly, had to enable emulation in Alcohol?
I am well aware of this rumour of collusion between Plextor and
Macrovision around their 1210 drive. We (cdfreaks) checked it directly
at the source and it was as expected completely false, and all recent
Plextor drives prove it.
Note also that Plextor is the only manufacturer which offers hardware
anti-protection features (single session, hide cd-r) and which sells
a drive able to make working backups of Securom (Premium).
Speaking of track quality, how many CDs have tracks with quality less than 100%? I am just curious. I have ripped many brand new, un-scratched, out-of-the-sealed-box CDS with NO fingerprints and most of them have at least one track with quailty of 99.99%. The number of such tracks varies from 1 to as much as 7 per disc. The average is about two tracks per disc. I am worried about the physical condition of the discs. Since they are brand new, I expect them to be perfect, but as you can see, it is not so.
How many people have seen this happening while you rip discs? Please let me know, thanks.
If you still intrested
This also depends on disk manufacturer, from my expirience. For example, I have Roxette and Blackmore's night series (3 disks) bought in Germany and even having scratches more than 50% tracks are ripped with EAC showing 100% quality.
But some local music disks (even new, w/o any damage) can have no one (!) track that can be ripped to 100%...
Oh, when you look at such a disk, you can notice that it have irregular surface fill inside (like muar, or it looks like the liquid is inside)...
It's a bit old story, maybe it was a screwup of Plextor, I don't know...
Beating the old horse is not a sport I like. I don't know what made me bring out that old case...
Andavari, so you have two drives that give the same result, but both do caching.
This still seems to be not enough to proove that your extraction is "true", but thanks for your effort. What a pitty that you don't have digital out on your player!
I'd really like to know your accuraterip results when this will work.
This still seems to be not enough to proove that your extraction is "true", but thanks for your effort. What a pitty that you don't have digital out on your player!
I'd really like to know your accuraterip results when this will work.
CRC's also match on my old computer using Sony USB CD-RW Drive CRX100E/X, it's now very old but it's good at ripping -- even nearly impossible discs.
I'm going to have to wait for the AccurateRip database to include the cd I have since it isn't listed in the database on the website, mine is too scratched to submit the ripping status from.
pio, what do you mean you can *see* the tracks on a cd!!?? i don't see a damn thing! are you a cyborg? ...
You can go to this thread to read more:
http://www.hydrogenaudio.org/show.php/showtopic/21628 (http://www.hydrogenaudio.org/show.php/showtopic/21628)
But the basic jist is that I have noticed that track quality of 98% and below, EVEN if NO errors are reported, does not tend to be error free when compared with the accuraterip results
Eli, in the thread you mentioned there seems to be no answer for PhileasFogg's question.
PhileasFogg, take a close look at a CD with silent (100% silence reported by EAC) gaps that are at least two seconds long. Use a bright light, turn the CD in your hand so that the light is reflected differently. You'll see what Pio and I were talking about.
But the basic jist is that I have noticed that track quality of 98% and below, EVEN if NO errors are reported, does not tend to be error free when compared with the accuraterip results
(Eli)
You're right to not trust less than 99,99% track quality. If this is the case, you must do the things I've mentioned in this (http://www.hydrogenaudio.org/forums/index.php?showtopic=21004&view=findpost&p=209402) post.
pio, what do you mean you can *see* the tracks on a cd!!?? i don't see a damn thing! are you a cyborg? ...
Look at these :
http://perso.numericable.fr/laguill2/pictures/cdtracks.jpg (http://perso.numericable.fr/laguill2/pictures/cdtracks.jpg) (163 kB)
You can see them in a dark room with ponctual bright lightning (mini halogen lamps are good) in your back. They appear mostly in the golden shining part.
I never paid that much attention to the pressed audio disc's before and had noticed
some data discs having a very irregular pattern almost like a normal ring that was wet
and smeared around.
In grabbing several all are slightly different but none so far like yours 'Pio'. Yours looks like a vinyl disc, having different distances between the rings designating the varying
track lengths.
So far the handful I looked at either have many rings (maybe 50+) or dont exhibit any type of rings but have that distinguishing pattern as like when writing only part of cdr.
Have a question back on the original subject about those that talk about 'not trusting'
or 'not being error free' if NOT 100%.
Are you guys just being 'overly sensitive' (Not being a bad thing) and noticed a couple
bytes differening at the very beginning or very end of a track possibly relating to
the 'synchronize between tracks' causing it or have you heard a definite audible error when that happens?
Just want to be sure when I relay an answer to this question since usually quoting
the EAC faq's and from the very informal or unscientific tests ive done never noticed
a problem that I'm stating for the most part truthful information.
I've never heard anything wrong and I am overly sensitive...
However, investigating more, I finally found a bug in my drive : it silences the end of the last track, and this causes the quality to be inferior to 100 % (not sure if it reports an error at the end).
For the track pattern on CDs, it is quite difficult to see. Since in my room, with a signle light on the ceinling, I couldn't see anything, I lit a lamp in next room, and positionned the CD in the dark, facing the distant light in the next room.
Then, when I can see the lamp reflecting in the CD, there are no rainbows.
Turning the CD downwards, some rainbows appear, aligned with the lamp reflection, that is no more visible.
On both sides of this rainbow, there is a dark zone, then a golden one. The track separation is most visible in these zones, between the lamp reflection and the begining of the main rainbow.
We can also notice in the above picture that before and after the data zone, there are some different zones. They feature dull rainbows, but parallel to the groove, unlike the bright main one that is perpendicular. It's these rainbows that can be seen between tracks. These zones all feature null data (before audio, between tracks, after audio).
teleguise,
Let's say you have the following result for a track:
Track quality about 99,98% or so, CRC=OK and, during the extraction, only the first row of the red lights lighted up. Then I'd accept this extraction as being "true" and I
am overly sensitive.
I only had one single case out of a thousand or so when CRC was OK and there were two single samples different from the probably "true" extraction. And no, I can't here a single wrong sample, and I'm sure anybody else in the world also can't.
There were never audible errors, unless EAC did report a 'sync error' or 'read error'.
But one exception: I have an old drive, that creates audible clicks at the end of the tracks and EAC doesn't notice it. But the reason is definitely the drive. So, if you've once compared your extraction with the extraction from a second drive and they were identical, and you have the conditions I've mentioned above, you really don't need to worry.
as for the CD data surface issue:
These zones all feature null data (before audio, between tracks, after audio).
(Pio)
I know I've suggested this 'null data' explanation, but have you read what I've written in this (http://www.hydrogenaudio.org/forums/index.php?showtopic=21004&view=findpost&p=207005) post, Pio ?
One isolated wrong sample can be audible if it was not detected and corrected by the drive, but this is a nearly SF scenario. First you'd have to get one isolated error (with at least 3 wrong samples. It should not possible to get less than 3 wrong samples at once), which is rare, then, it would have to be an undetected, thus uncorrected error (speaking about the internal processes of the drive) in order to be audible, which is even more unlikely (statistically 1 % of these rare cases). Last, internal error detection/correction failure seem to occur in extreme conditions (the lens can't follow the groove anymore, for example), which implies sourrounding errors.
So yes an isolated sample can make a lot of noise, but you won't find just one on a CD because of a read error (there will be many more with it )
For the spaces between the tracks, even if it is not perfect silence, it can be a dithered silence oscillating between two values. The presence of two kind of codes instead of 65536 must be able to produce visible marks.
In fact samples are divided into bytes. So it is even possible that parts with peak level below -6 db are visible, since they are made with 50 % of null bytes.
So are you saying that if the track quality falls below say due to it going
two the 2nd or 3rd row that you don't consider this 'true' anymore?
I gather each row is computed individually & doesn't each row (up to 5) simply mean a
re-read of 16x's and that EAC doesn't consider it correct unless (1/2) 8 of them are
identical?
Still confused how the term 'up to 82 times' multiplies out, does the very first read + re-read account for the 2 additional read times and not inclusive when going into error recover? Hence 2+(16x5)=82
PIO you know way to much
I gather each row is computed individually & doesn't each row (up to 5) simply mean a
re-read of 16x's and that EAC doesn't consider it correct unless (1/2) 8 of them are
identical?
A fas as I know, that's right. This 82 figure must be a "marketing trick"
teleguise,
what you said seems to be right.
It doesn't matter that much if the track quality is 99,9% or 99,8%. The important thing is CRC=OK, since this is very unlikely (unless the rip is 'true'). My experience is that whenever EAC needs more than the first row, it mostly needs all rows, and if no error is reported, then most likely CRC=#. But sometimes EAC needs the first row of error correction at the end of some tracks (perfect CDs, perfect settings etc.). This seems to be a bug and leads to track quality= less than 100,0%, but it causes no further problems or errors (and of course it isn't audible).
Pio,
my point of view was one or very few samples which are few sample values out of range, just to make it sure.